432 ALASKAN GLACIER STUDIES record, the level going up over 21 feet above normal winter level, and falling rapidly again. Johnson states that the water came from beneath Miles Glacier at the point where the ice mass floated out 5 months later, and that "from the time of the break the river continued to rise for four days, was stationary one day, and then receded at about the same rapidity as the rise. The river at the bridge crossing . . . (before this flood) had a cross-section area of about ten thousand square feet and an average velocity of two miles per hour, but during the highest stage of the river the cross-section area was about forty thousand square feet, and average velocity ten miles per hour. . . . Considerable ice fell from the face of the glacier at the point of the break during the flow, but not until the middle of July did any great amount break away—at which time a section at least one half mile square, caved, floated out and completely filled the lake, which covers approximately five square miles." The climatic record kept by the railway engineers at Miles Glacier bridge, shows no rainfall at or before the time of the flood and only the normal increase of snowfall for that season of the year. There were fluctuations in temperature before the flood began on February 9, but from the middle of December until two weeks after the flood the maximum temperature was not above the freezing point. The wind velocity did not increase abnormally immediately before the flood, though on the day when the water was highest the wind blew 48 to 84 miles an hour from the northeast. A high northeast wind in summer would drive the water from the place where the break occurred to the outlet of the lake but it could not have caused the initiation of the flood in February, for the lake was then frozen over, and the flood could not have kept up unless water was supplied from outside. The meteorological conditions, therefore, seem to have had little, if anything, to do with the flood. Our own view of the detaching of the ice mass in 1909 is that it was not necessarily related to the advance or retreat of the glacier at all, but to undermining during the February flood which may have been caused by the draining of a marginal lake higher up Miles Glacier. A similar flood on August 16, 1912, perhaps from the draining of a marginal lake, swept down the Copper River from Miles Glacier. It raised the water level 12 feet at the railway bridge east of Childs Glacier and, 20 miles farther south, swept away 1600 feet of railway trestle east of Flag Point, drowning a repair crew foreman. A second event of similar character occurred during the summer of 1910. At the time of our visit in 1909 and up to June, 1910, there was a stagnant, projecting edge of Miles Glacier which clung to the mountain side at the southern end of the ice cliff and merged into the glacial deposits there. Sometime between June 10 and August 15, 1910, a section of the glacier, measuring over an eighth of a mile on each side, floated away, leaving the outer stagnant ice as a detached block, 800 by 1200 feet, and extending up the mountain slope to a height of 175 to 200 feet above the lake. The whole of this stagnant block was thickly covered with ablation moraine, and it was riven by great cracks. To the southward it merged into slumping gravel deposits. Since the southern ice edge was still advancing after this separation took place, one possible method by which the separation may have been effected is the shove of the advancing glacier against the stagnant border which was clinging to the mountain side, thus placing a portion of it in such unstable equilibrium that it broke up. In June, 1911, Miles Glacier showed no appreciable changes since the previous year.